[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-DamRsn--NeuralNote":3,"tool-DamRsn--NeuralNote":64},[4,17,27,35,43,56],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":16},3808,"stable-diffusion-webui","AUTOMATIC1111\u002Fstable-diffusion-webui","stable-diffusion-webui 是一个基于 Gradio 构建的网页版操作界面，旨在让用户能够轻松地在本地运行和使用强大的 Stable Diffusion 图像生成模型。它解决了原始模型依赖命令行、操作门槛高且功能分散的痛点，将复杂的 AI 绘图流程整合进一个直观易用的图形化平台。\n\n无论是希望快速上手的普通创作者、需要精细控制画面细节的设计师，还是想要深入探索模型潜力的开发者与研究人员，都能从中获益。其核心亮点在于极高的功能丰富度：不仅支持文生图、图生图、局部重绘（Inpainting）和外绘（Outpainting）等基础模式，还独创了注意力机制调整、提示词矩阵、负向提示词以及“高清修复”等高级功能。此外，它内置了 GFPGAN 和 CodeFormer 等人脸修复工具，支持多种神经网络放大算法，并允许用户通过插件系统无限扩展能力。即使是显存有限的设备，stable-diffusion-webui 也提供了相应的优化选项，让高质量的 AI 艺术创作变得触手可及。",162132,3,"2026-04-05T11:01:52",[13,14,15],"开发框架","图像","Agent","ready",{"id":18,"name":19,"github_repo":20,"description_zh":21,"stars":22,"difficulty_score":23,"last_commit_at":24,"category_tags":25,"status":16},1381,"everything-claude-code","affaan-m\u002Feverything-claude-code","everything-claude-code 是一套专为 AI 编程助手（如 Claude Code、Codex、Cursor 等）打造的高性能优化系统。它不仅仅是一组配置文件，而是一个经过长期实战打磨的完整框架，旨在解决 AI 代理在实际开发中面临的效率低下、记忆丢失、安全隐患及缺乏持续学习能力等核心痛点。\n\n通过引入技能模块化、直觉增强、记忆持久化机制以及内置的安全扫描功能，everything-claude-code 能显著提升 AI 在复杂任务中的表现，帮助开发者构建更稳定、更智能的生产级 AI 代理。其独特的“研究优先”开发理念和针对 Token 消耗的优化策略，使得模型响应更快、成本更低，同时有效防御潜在的攻击向量。\n\n这套工具特别适合软件开发者、AI 研究人员以及希望深度定制 AI 工作流的技术团队使用。无论您是在构建大型代码库，还是需要 AI 协助进行安全审计与自动化测试，everything-claude-code 都能提供强大的底层支持。作为一个曾荣获 Anthropic 黑客大奖的开源项目，它融合了多语言支持与丰富的实战钩子（hooks），让 AI 真正成长为懂上",138956,2,"2026-04-05T11:33:21",[13,15,26],"语言模型",{"id":28,"name":29,"github_repo":30,"description_zh":31,"stars":32,"difficulty_score":23,"last_commit_at":33,"category_tags":34,"status":16},2271,"ComfyUI","Comfy-Org\u002FComfyUI","ComfyUI 是一款功能强大且高度模块化的视觉 AI 引擎，专为设计和执行复杂的 Stable Diffusion 图像生成流程而打造。它摒弃了传统的代码编写模式，采用直观的节点式流程图界面，让用户通过连接不同的功能模块即可构建个性化的生成管线。\n\n这一设计巧妙解决了高级 AI 绘图工作流配置复杂、灵活性不足的痛点。用户无需具备编程背景，也能自由组合模型、调整参数并实时预览效果，轻松实现从基础文生图到多步骤高清修复等各类复杂任务。ComfyUI 拥有极佳的兼容性，不仅支持 Windows、macOS 和 Linux 全平台，还广泛适配 NVIDIA、AMD、Intel 及苹果 Silicon 等多种硬件架构，并率先支持 SDXL、Flux、SD3 等前沿模型。\n\n无论是希望深入探索算法潜力的研究人员和开发者，还是追求极致创作自由度的设计师与资深 AI 绘画爱好者，ComfyUI 都能提供强大的支持。其独特的模块化架构允许社区不断扩展新功能，使其成为当前最灵活、生态最丰富的开源扩散模型工具之一，帮助用户将创意高效转化为现实。",107662,"2026-04-03T11:11:01",[13,14,15],{"id":36,"name":37,"github_repo":38,"description_zh":39,"stars":40,"difficulty_score":23,"last_commit_at":41,"category_tags":42,"status":16},3704,"NextChat","ChatGPTNextWeb\u002FNextChat","NextChat 是一款轻量且极速的 AI 助手，旨在为用户提供流畅、跨平台的大模型交互体验。它完美解决了用户在多设备间切换时难以保持对话连续性，以及面对众多 AI 模型不知如何统一管理的痛点。无论是日常办公、学习辅助还是创意激发，NextChat 都能让用户随时随地通过网页、iOS、Android、Windows、MacOS 或 Linux 端无缝接入智能服务。\n\n这款工具非常适合普通用户、学生、职场人士以及需要私有化部署的企业团队使用。对于开发者而言，它也提供了便捷的自托管方案，支持一键部署到 Vercel 或 Zeabur 等平台。\n\nNextChat 的核心亮点在于其广泛的模型兼容性，原生支持 Claude、DeepSeek、GPT-4 及 Gemini Pro 等主流大模型，让用户在一个界面即可自由切换不同 AI 能力。此外，它还率先支持 MCP（Model Context Protocol）协议，增强了上下文处理能力。针对企业用户，NextChat 提供专业版解决方案，具备品牌定制、细粒度权限控制、内部知识库整合及安全审计等功能，满足公司对数据隐私和个性化管理的高标准要求。",87618,"2026-04-05T07:20:52",[13,26],{"id":44,"name":45,"github_repo":46,"description_zh":47,"stars":48,"difficulty_score":23,"last_commit_at":49,"category_tags":50,"status":16},2268,"ML-For-Beginners","microsoft\u002FML-For-Beginners","ML-For-Beginners 是由微软推出的一套系统化机器学习入门课程，旨在帮助零基础用户轻松掌握经典机器学习知识。这套课程将学习路径规划为 12 周，包含 26 节精炼课程和 52 道配套测验，内容涵盖从基础概念到实际应用的完整流程，有效解决了初学者面对庞大知识体系时无从下手、缺乏结构化指导的痛点。\n\n无论是希望转型的开发者、需要补充算法背景的研究人员，还是对人工智能充满好奇的普通爱好者，都能从中受益。课程不仅提供了清晰的理论讲解，还强调动手实践，让用户在循序渐进中建立扎实的技能基础。其独特的亮点在于强大的多语言支持，通过自动化机制提供了包括简体中文在内的 50 多种语言版本，极大地降低了全球不同背景用户的学习门槛。此外，项目采用开源协作模式，社区活跃且内容持续更新，确保学习者能获取前沿且准确的技术资讯。如果你正寻找一条清晰、友好且专业的机器学习入门之路，ML-For-Beginners 将是理想的起点。",84991,"2026-04-05T10:45:23",[14,51,52,53,15,54,26,13,55],"数据工具","视频","插件","其他","音频",{"id":57,"name":58,"github_repo":59,"description_zh":60,"stars":61,"difficulty_score":10,"last_commit_at":62,"category_tags":63,"status":16},3128,"ragflow","infiniflow\u002Fragflow","RAGFlow 是一款领先的开源检索增强生成（RAG）引擎，旨在为大语言模型构建更精准、可靠的上下文层。它巧妙地将前沿的 RAG 技术与智能体（Agent）能力相结合，不仅支持从各类文档中高效提取知识，还能让模型基于这些知识进行逻辑推理和任务执行。\n\n在大模型应用中，幻觉问题和知识滞后是常见痛点。RAGFlow 通过深度解析复杂文档结构（如表格、图表及混合排版），显著提升了信息检索的准确度，从而有效减少模型“胡编乱造”的现象，确保回答既有据可依又具备时效性。其内置的智能体机制更进一步，使系统不仅能回答问题，还能自主规划步骤解决复杂问题。\n\n这款工具特别适合开发者、企业技术团队以及 AI 研究人员使用。无论是希望快速搭建私有知识库问答系统，还是致力于探索大模型在垂直领域落地的创新者，都能从中受益。RAGFlow 提供了可视化的工作流编排界面和灵活的 API 接口，既降低了非算法背景用户的上手门槛，也满足了专业开发者对系统深度定制的需求。作为基于 Apache 2.0 协议开源的项目，它正成为连接通用大模型与行业专有知识之间的重要桥梁。",77062,"2026-04-04T04:44:48",[15,14,13,26,54],{"id":65,"github_repo":66,"name":67,"description_en":68,"description_zh":69,"ai_summary_zh":69,"readme_en":70,"readme_zh":71,"quickstart_zh":72,"use_case_zh":73,"hero_image_url":74,"owner_login":75,"owner_name":76,"owner_avatar_url":77,"owner_bio":78,"owner_company":78,"owner_location":78,"owner_email":78,"owner_twitter":78,"owner_website":78,"owner_url":79,"languages":80,"stars":109,"forks":110,"last_commit_at":111,"license":112,"difficulty_score":23,"env_os":113,"env_gpu":114,"env_ram":114,"env_deps":115,"category_tags":124,"github_topics":125,"view_count":132,"oss_zip_url":78,"oss_zip_packed_at":78,"status":16,"created_at":133,"updated_at":134,"faqs":135,"releases":165},208,"DamRsn\u002FNeuralNote","NeuralNote","Audio Plugin for Audio to MIDI transcription using deep learning.","NeuralNote 是一款专为音乐创作设计的音频插件，利用深度学习技术将音频信号实时转换为 MIDI 音符。NeuralNote 有效解决了音乐制作中手动记谱耗时费力的痛点，无论是人声演唱还是多种乐器演奏，包括复杂的复调音乐，都能进行准确识别与转录。\n\nNeuralNote 非常适合音乐制作人、作曲家以及需要在数字音频工作站（DAW）中快速提取旋律素材的用户。核心算法源自 Spotify 的 basic-pitch 模型，并通过 RTNeural 和 ONNXRuntime 进行了深度优化，实现了轻量级且高速的转录性能。\n\n在使用过程中，用户可以实时调整参数并监听转录效果，直接在插件内完成音高缩放和时间量化。满意后，只需简单拖拽即可将 MIDI 导出至工程轨道。NeuralNote 支持 Windows、macOS 和 Linux 系统，提供 VST3、AU 及独立应用多种版本，让 AI 辅助创作变得简单高效。","# NeuralNote \u003Cimg style=\"float: right;\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FDamRsn_NeuralNote_readme_8c44f98c276a.png\" width=\"100\" \u002F>\n\nNeuralNote is the audio plugin that brings **state-of-the-art Audio to MIDI conversion** into\nyour favorite Digital Audio Workstation.\n\n- Works with any tonal instrument (voice included)\n- Supports polyphonic transcription\n- Supports pitch bend detection\n- Lightweight and very fast transcription\n- Allows to adjust the parameters while listening to the transcription\n- Allows to scale and time quantize transcribed MIDI directly in the plugin\n\n## Install NeuralNote\n\nDownload the latest release for your platform [here](https:\u002F\u002Fgithub.com\u002FDamRsn\u002FNeuralNote\u002Freleases) (Windows, macOS (\nUniversal) and Linux supported)!\n\nInstallers are available for both Windows and Mac, including Standalone, VST3, and AU (Mac only) versions. The\ninstallers allow users to select which format(s) they want to install. On macOS, the code is signed, while on Windows,\nit is not. This means you may need to take a few additional steps to use NeuralNote on Windows.\n\nFor Linux, raw binaries are provided for VST3 and Standalone. You can install them by copying the files to the\nappropriate locations.\n\n## Usage\n\n![UI](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FDamRsn_NeuralNote_readme_7c2bd96dd448.png)\n\nNeuralNote comes as a simple AudioFX plugin (VST3\u002FAU\u002FStandalone app) to be applied on the track to transcribe.\n\nThe workflow is very simple:\n\n- Gather some audio\n    - Click record. Works when recording for real or when playing the track in a DAW.\n    - Or drop an audio file on the plugin. (.wav, .aiff, .flac, .mp3 and .ogg (vorbis) supported)\n- The MIDI transcription instantly appears in the piano roll section.\n- Listen to the result by clicking the play button.\n    - Play with the different settings to adjust the transcription, even while listening to it\n    - Individually adjust the level of the source audio and of the synthesized transcription\n- Once you're satisfied, export the MIDI transcription with a simple drag and drop from the plugin to a MIDI track.\n\n**Watch our presentation video for the Neural Audio Plugin\ncompetition [here](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=6_MC0_aG_DQ)**.\n\nNeuralNote uses internally the model from Spotify's [basic-pitch](https:\u002F\u002Fgithub.com\u002Fspotify\u002Fbasic-pitch). See\ntheir [blogpost](https:\u002F\u002Fengineering.atspotify.com\u002F2022\u002F06\u002Fmeet-basic-pitch\u002F)\nand [paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2203.09893) for more information. In NeuralNote, basic-pitch is run\nusing [RTNeural](https:\u002F\u002Fgithub.com\u002Fjatinchowdhury18\u002FRTNeural) for the CNN part\nand [ONNXRuntime](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fonnxruntime) for the feature part (Constant-Q transform calculation +\nHarmonic Stacking).\nAs part of this project, [we contributed to RTNeural](https:\u002F\u002Fgithub.com\u002Fjatinchowdhury18\u002FRTNeural\u002Fpull\u002F89) to add 2D\nconvolution support.\n\n## Build from source\n\nRequirements are: `git`, `cmake`, and your OS's preferred compiler suite.\n\nUse this when cloning:\n\n```\ngit clone --recurse-submodules --shallow-submodules https:\u002F\u002Fgithub.com\u002FDamRsn\u002FNeuralNote\n ```\n\nThe following OS-specific build scripts have to be executed at least once before being able to use the project as a\nnormal CMake project. The script downloads onnxruntime static library (that we created\nwith [ort-builder](https:\u002F\u002Fgithub.com\u002Folilarkin\u002Fort-builder)) before calling CMake.\n\n#### macOS\n\n```\n$ .\u002Fbuild.sh\n```\n\n#### Windows\n\nDue to [a known issue](https:\u002F\u002Fgithub.com\u002FDamRsn\u002FNeuralNote\u002Fissues\u002F21), if you're not using Visual Studio 2022 (MSVC\nversion: 19.35.x, check `cl` output), then you'll need to manually build onnxruntime.lib like so:\n\n1. Ensure you have Python installed; if not, download at https:\u002F\u002Fwww.python.org\u002Fdownloads\u002Fwindows\u002F (this does not\n   currently work with Python 3.11, prefer Python 3.10).\n\n2. Execute each of the following lines in a command prompt:\n\n```\ngit clone --depth 1 --recurse-submodules --shallow-submodules https:\u002F\u002Fgithub.com\u002Ftiborvass\u002Flibonnxruntime-neuralnote ThirdParty\\onnxruntime\ncd ThirdParty\\onnxruntime\npython3 -m venv venv\n.\\venv\\Scripts\\activate.bat\npip install -r requirements.txt\n.\\convert-model-to-ort.bat model.onnx\n.\\build-win.bat model.required_operators_and_types.with_runtime_opt.config\ncopy model.with_runtime_opt.ort ..\\..\\Lib\\ModelData\\features_model.ort\ncd ..\\..\n```\n\nNow you can get back to building NeuralNote as follows:\n\n```\n> .\\build.bat\n```\n\n#### IDEs\n\nOnce the build script has been executed at least once, you can load this project in your favorite IDE\n(CLion\u002FVisual Studio\u002FVSCode\u002Fetc) and click 'build' for one of the targets.\n\n## Reuse code from NeuralNote’s transcription engine\n\nAll the code to perform the transcription is in `Lib\u002FModel` and all the model weights are in `Lib\u002FModelData\u002F`. Feel free\nto use only this part of the code in your own project! We'll try to isolate it more from the rest of the repo in the\nfuture and make it a library.\n\nThe code to generate the files in `Lib\u002FModelData\u002F` is not currently available as it required a lot of manual operations.\nBut here's a description of the process we followed to create those files:\n\n- `features_model.onnx` was generated by converting a keras model containing only the CQT + Harmonic Stacking part of\n  the full basic-pitch graph using `tf2onnx` (with manually added weights for batch normalization).\n- the `.json` files containing the weights of the basic-pitch cnn were generated from the tensorflow-js model available\n  in the [basic-pitch-ts repository](https:\u002F\u002Fgithub.com\u002Fspotify\u002Fbasic-pitch-ts), then converted to onnx with `tf2onnx`.\n  Finally, the weights were gathered manually to `.npy` thanks to [Netron](https:\u002F\u002Fnetron.app\u002F) and finally applied to a\n  split keras model created with [basic-pitch](https:\u002F\u002Fgithub.com\u002Fspotify\u002Fbasic-pitch) code.\n\nThe original basic-pitch CNN was split in 4 sequential models wired together, so they can be run with RTNeural.\n\n## Bug reports and feature requests\n\nIf you have any request\u002Fsuggestion concerning the plugin or encounter a bug, please file a GitHub issue.\n\n## Contributing\n\nContributions are most welcome! If you want to add some features to the plugin or simply improve the documentation,\nplease open a PR!\n\n## License\n\nNeuralNote software and code is published under the Apache-2.0 license. See the [license file](LICENSE).\n\n#### Third Party libraries used and license\n\nHere's a list of all the third party libraries used in NeuralNote and the license under which they are used.\n\n- [JUCE](https:\u002F\u002Fjuce.com\u002F) (JUCE Starter)\n- [RTNeural](https:\u002F\u002Fgithub.com\u002Fjatinchowdhury18\u002FRTNeural) (BSD-3-Clause license)\n- [ONNXRuntime](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fonnxruntime) (MIT License)\n- [ort-builder](https:\u002F\u002Fgithub.com\u002Folilarkin\u002Fort-builder) (MIT License)\n- [basic-pitch](https:\u002F\u002Fgithub.com\u002Fspotify\u002Fbasic-pitch) (Apache-2.0 license)\n- [basic-pitch-ts](https:\u002F\u002Fgithub.com\u002Fspotify\u002Fbasic-pitch-ts) (Apache-2.0 license)\n- [minimp3](https:\u002F\u002Fgithub.com\u002Flieff\u002Fminimp3) (CC0-1.0 license)\n\n## Could NeuralNote transcribe audio in real-time?\n\nUnfortunately no and this for a few reasons:\n\n- Basic Pitch uses the Constant-Q transform (CQT) as input feature. The CQT requires really long audio chunks (> 1s) to\n  get amplitudes for the lowest frequency bins. This makes the latency too high to have real-time transcription.\n- The basic pitch CNN has an additional latency of approximately 120ms.\n- The note events creation algorithm processes the posteriorgrams backward (from future to past) and is hence\n  non-causal.\n\nBut if you have ideas please share!\n\n## Credits\n\nNeuralNote was developed by [Damien Ronssin](https:\u002F\u002Fgithub.com\u002FDamRsn) and [Tibor Vass](https:\u002F\u002Fgithub.com\u002Ftiborvass).\nThe plugin user interface was designed by Perrine Morel.\n\n#### Contributors\n\nMany thanks to the contributors!\n\n- [jatinchowdhury18](https:\u002F\u002Fgithub.com\u002Fjatinchowdhury18): File browser.\n- [trirpi](https:\u002F\u002Fgithub.com\u002Ftrirpi)\n    - More scale options in `SCALE QUANTIZE`.\n    - Horizontal zoom for the audio waveform and the piano roll.\n- [polygon](https:\u002F\u002Fgithub.com\u002Fpolygon) and [SamuMazzi](https:\u002F\u002Fgithub.com\u002FSamuMazzi): Linux support.","# NeuralNote \u003Cimg style=\"float: right;\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FDamRsn_NeuralNote_readme_8c44f98c276a.png\" width=\"100\" \u002F>\n\nNeuralNote 是一款音频插件，将**最先进的音频转 MIDI 转换（Audio to MIDI conversion）**功能带入您喜爱的数字音频工作站（Digital Audio Workstation）。\n\n- 适用于任何有音高乐器（tonal instrument，包括人声）\n- 支持复调转录（polyphonic transcription）\n- 支持弯音检测（pitch bend detection）\n- 轻量且非常快速的转录\n- 允许在聆听转录结果时调整参数\n- 允许直接在插件中对转录的 MIDI 进行缩放和时间量化（time quantize）\n\n## 安装 NeuralNote\n\n在此处下载适用于您平台的最新版本 [here](https:\u002F\u002Fgithub.com\u002FDamRsn\u002FNeuralNote\u002Freleases)（支持 Windows、macOS (Universal) 和 Linux）！\n\nWindows 和 Mac 均有安装程序可用，包括独立应用（Standalone）、VST3 和 AU（仅 Mac）版本。安装程序允许用户选择想要安装的格式。在 macOS 上，代码已签名，而在 Windows 上则未签名。这意味着您在 Windows 上使用 NeuralNote 可能需要采取一些额外步骤。\n\n对于 Linux，提供了 VST3 和独立应用（Standalone）的原始二进制文件。您可以通过将文件复制到适当的位置来安装它们。\n\n## 使用方法\n\n![UI](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FDamRsn_NeuralNote_readme_7c2bd96dd448.png)\n\nNeuralNote 作为一个简单的 AudioFX 插件（VST3\u002FAU\u002F独立应用）提供，应用于要转录的轨道。\n\n工作流程非常简单：\n\n- 获取一些音频\n    - 点击录音。适用于真实录音或在数字音频工作站（DAW）中播放轨道时。\n    - 或将音频文件拖放到插件上。（支持 .wav, .aiff, .flac, .mp3 和 .ogg (vorbis)）\n- MIDI 转录结果会立即出现在钢琴卷帘（piano roll）部分。\n- 点击播放按钮聆听结果。\n    - 尝试不同的设置以调整转录，甚至在聆听时也可调整\n    - 单独调整源音频和合成转录的音量级别\n- 满意后，只需将 MIDI 转录结果从插件拖放到 MIDI 轨道即可导出。\n\n**在此处观看我们为 Neural Audio Plugin 竞赛制作的演示视频 [here](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=6_MC0_aG_DQ)**。\n\nNeuralNote 内部使用 Spotify 的 [basic-pitch](https:\u002F\u002Fgithub.com\u002Fspotify\u002Fbasic-pitch) 模型。请参阅他们的 [博客文章](https:\u002F\u002Fengineering.atspotify.com\u002F2022\u002F06\u002Fmeet-basic-pitch\u002F) 和 [论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2203.09893) 以获取更多信息。在 NeuralNote 中，basic-pitch 使用 [RTNeural](https:\u002F\u002Fgithub.com\u002Fjatinchowdhury18\u002FRTNeural) 运行卷积神经网络（CNN）部分，使用 [ONNXRuntime](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fonnxruntime) 运行特征部分（恒定 Q 变换（Constant-Q transform）计算 + 谐波堆叠（Harmonic Stacking））。作为本项目的一部分，[我们为 RTNeural 做出了贡献](https:\u002F\u002Fgithub.com\u002Fjatinchowdhury18\u002FRTNeural\u002Fpull\u002F89) 以添加 2D 卷积支持。\n\n## 从源代码构建\n\n要求：`git`、`cmake` 和您操作系统首选的编译器套件。\n\n克隆时使用此命令：\n\n```\ngit clone --recurse-submodules --shallow-submodules https:\u002F\u002Fgithub.com\u002FDamRsn\u002FNeuralNote\n ```\n\n以下特定于操作系统的构建脚本在能够将项目作为普通 CMake 项目使用之前必须至少执行一次。该脚本在调用 CMake 之前下载 onnxruntime 静态库（我们使用 [ort-builder](https:\u002F\u002Fgithub.com\u002Folilarkin\u002Fort-builder) 创建）。\n\n#### macOS\n\n```\n$ .\u002Fbuild.sh\n```\n\n#### Windows\n\n由于 [一个已知问题](https:\u002F\u002Fgithub.com\u002FDamRsn\u002FNeuralNote\u002Fissues\u002F21)，如果您不使用 Visual Studio 2022（MSVC 版本：19.35.x，检查 `cl` 输出），那么您需要手动构建 onnxruntime.lib，如下所示：\n\n1. 确保您已安装 Python；如果没有，请在 https:\u002F\u002Fwww.python.org\u002Fdownloads\u002Fwindows\u002F 下载（目前不适用于 Python 3.11，建议使用 Python 3.10）。\n\n2. 在命令提示符中执行以下每一行：\n\n```\ngit clone --depth 1 --recurse-submodules --shallow-submodules https:\u002F\u002Fgithub.com\u002Ftiborvass\u002Flibonnxruntime-neuralnote ThirdParty\\onnxruntime\ncd ThirdParty\\onnxruntime\npython3 -m venv venv\n.\\venv\\Scripts\\activate.bat\npip install -r requirements.txt\n.\\convert-model-to-ort.bat model.onnx\n.\\build-win.bat model.required_operators_and_types.with_runtime_opt.config\ncopy model.with_runtime_opt.ort ..\\..\\Lib\\ModelData\\features_model.ort\ncd ..\\..\n```\n\n现在您可以回到构建 NeuralNote，如下所示：\n\n```\n> .\\build.bat\n```\n\n#### 集成开发环境（IDEs）\n\n一旦构建脚本至少执行过一次，您就可以在您喜欢的 IDE（CLion\u002FVisual Studio\u002FVSCode 等）中加载此项目，并为其中一个目标点击“构建”。\n\n## 复用 NeuralNote 转录引擎的代码\n\n所有执行转录的代码都在 `Lib\u002FModel` 中，所有模型权重都在 `Lib\u002FModelData\u002F` 中。欢迎在您自己的项目中仅使用这部分代码！我们将尝试在未来将其与仓库的其余部分更加隔离，并将其制作成一个库。\n\n生成 `Lib\u002FModelData\u002F` 中文件的代码目前不可用，因为它需要大量手动操作。但以下是我们创建这些文件的过程描述：\n\n- `features_model.onnx` 是通过使用 `tf2onnx` 转换一个仅包含完整 basic-pitch 图谱中 CQT + 谐波堆叠（Harmonic Stacking）部分的 keras 模型生成的（手动添加了批归一化（batch normalization）的权重）。\n- 包含 basic-pitch cnn 权重的 `.json` 文件是从 [basic-pitch-ts 仓库](https:\u002F\u002Fgithub.com\u002Fspotify\u002Fbasic-pitch-ts) 中可用的 tensorflow-js 模型生成的，然后使用 `tf2onnx` 转换为 onnx。最后，借助 [Netron](https:\u002F\u002Fnetron.app\u002F) 将权重手动收集到 `.npy`，最后应用到使用 [basic-pitch](https:\u002F\u002Fgithub.com\u002Fspotify\u002Fbasic-pitch) 代码创建的拆分 keras 模型。\n\n原始的 basic-pitch CNN 被拆分为 4 个顺序连接的模型，以便它们可以与 RTNeural 一起运行。\n\n## 错误报告和功能请求\n\n如果您有任何关于插件的请求\u002F建议或遇到错误，请提交 GitHub issue。\n\n## 贡献\n\n非常欢迎贡献！如果您想为插件添加一些功能或 simply 改进文档，请开启一个拉取请求（PR）！\n\n## 许可证\n\nNeuralNote 软件和代码根据 Apache-2.0 许可证发布。请参阅 [许可证文件](LICENSE)。\n\n#### 使用的第三方库及许可证\n\n以下是 NeuralNote 中使用的所有第三方库及其使用许可证的列表。\n\n- [JUCE](https:\u002F\u002Fjuce.com\u002F) (JUCE Starter)\n- [RTNeural](https:\u002F\u002Fgithub.com\u002Fjatinchowdhury18\u002FRTNeural) (BSD-3-Clause license)\n- [ONNXRuntime](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fonnxruntime) (MIT License)\n- [ort-builder](https:\u002F\u002Fgithub.com\u002Folilarkin\u002Fort-builder) (MIT License)\n- [basic-pitch](https:\u002F\u002Fgithub.com\u002Fspotify\u002Fbasic-pitch) (Apache-2.0 license)\n- [basic-pitch-ts](https:\u002F\u002Fgithub.com\u002Fspotify\u002Fbasic-pitch-ts) (Apache-2.0 license)\n- [minimp3](https:\u002F\u002Fgithub.com\u002Flieff\u002Fminimp3) (CC0-1.0 license)\n\n## NeuralNote 可以实时转录音频吗？\n\n遗憾的是不行，原因有以下几点：\n\n- Basic Pitch 使用 Constant-Q transform (CQT)（恒 Q 变换）作为输入特征。CQT 需要非常长的音频块（> 1 秒）才能获取最低频率仓的振幅。这使得延迟过高，无法进行实时转录。\n- basic pitch CNN（卷积神经网络）还有大约 120ms 的额外延迟。\n- 音符事件创建算法向后处理 posteriorgrams（后验图）（从未来到过去），因此是 non-causal（非因果）的。\n\n但如果你有想法，请分享！\n\n## 致谢\n\nNeuralNote 由 [Damien Ronssin](https:\u002F\u002Fgithub.com\u002FDamRsn) 和 [Tibor Vass](https:\u002F\u002Fgithub.com\u002Ftiborvass) 开发。插件用户界面由 Perrine Morel 设计。\n\n#### 贡献者\n\n非常感谢各位贡献者！\n\n- [jatinchowdhury18](https:\u002F\u002Fgithub.com\u002Fjatinchowdhury18)：文件浏览器。\n- [trirpi](https:\u002F\u002Fgithub.com\u002Ftrirpi)\n    - `SCALE QUANTIZE` 中提供更多音阶选项。\n    - 音频波形和钢琴卷帘的水平缩放。\n- [polygon](https:\u002F\u002Fgithub.com\u002Fpolygon) 和 [SamuMazzi](https:\u002F\u002Fgithub.com\u002FSamuMazzi)：Linux 支持。","# NeuralNote 快速上手指南\n\nNeuralNote 是一款将音频转换为 MIDI 的音频插件，支持复调转录和音高弯曲检测，可集成至主流数字音频工作站（DAW）。\n\n## 环境准备\n\n**系统要求**\n- **操作系统**: Windows, macOS (Universal), Linux\n- **插件格式**: VST3, AU (仅 macOS), Standalone 应用\n- **硬件**: 支持任意音调乐器（含人声）\n\n**编译依赖（仅源码构建需要）**\n- `git`, `cmake`, 系统首选编译器套件\n- **Windows 额外要求**: 建议使用 Visual Studio 2022 (MSVC 版本 19.35.x)。若版本不符，需安装 Python 3.10（不支持 3.11）用于构建 onnxruntime。\n- **macOS 额外要求**: 代码已签名，可直接运行。\n- **Linux 额外要求**: 需手动复制二进制文件至合适位置。\n\n## 安装步骤\n\n### 方式一：使用预编译版本（推荐）\n1. 访问 [GitHub Releases](https:\u002F\u002Fgithub.com\u002FDamRsn\u002FNeuralNote\u002Freleases) 下载对应平台的最新安装包。\n2. **Windows\u002FmacOS**: 运行安装程序，选择需要安装的格式（Standalone, VST3, AU）。\n   > **注意**: Windows 版本代码未签名，首次运行可能需要额外授权步骤。\n3. **Linux**: 下载 VST3 和 Standalone 原始二进制文件，手动复制到插件扫描路径。\n\n### 方式二：源码编译\n1. 克隆项目仓库（包含子模块）：\n```bash\ngit clone --recurse-submodules --shallow-submodules https:\u002F\u002Fgithub.com\u002FDamRsn\u002FNeuralNote\n```\n2. 执行构建脚本（首次运行会下载 onnxruntime 静态库）：\n\n**macOS**\n```bash\n.\u002Fbuild.sh\n```\n\n**Windows**\n若 Visual Studio 版本不符合要求，需先手动构建 onnxruntime（确保 Python 3.10 环境）：\n```bash\ngit clone --depth 1 --recurse-submodules --shallow-submodules https:\u002F\u002Fgithub.com\u002Ftiborvass\u002Flibonnxruntime-neuralnote ThirdParty\\onnxruntime\ncd ThirdParty\\onnxruntime\npython3 -m venv venv\n.\\venv\\Scripts\\activate.bat\npip install -r requirements.txt\n.\\convert-model-to-ort.bat model.onnx\n.\\build-win.bat model.required_operators_and_types.with_runtime_opt.config\ncopy model.with_runtime_opt.ort ..\\..\\Lib\\ModelData\\features_model.ort\ncd ..\\..\n```\n随后执行主构建脚本：\n```bash\n.\\build.bat\n```\n3. 构建完成后，可在 IDE（CLion\u002FVisual Studio\u002FVSCode 等）中加载项目并编译目标。\n\n## 基本使用\n\n1. **加载插件**\n   在 DAW 中将 NeuralNote 作为 AudioFX 插件加载到需要转录的轨道，或直接打开 Standalone 应用。\n\n2. **输入音频**\n   - **录制**: 点击 Record 按钮（支持实时录制或 DAW 播放）。\n   - **导入**: 将音频文件拖拽至插件界面（支持 .wav, .aiff, .flac, .mp3, .ogg）。\n\n3. **调整与试听**\n   - 转录的 MIDI 会即时显示在钢琴卷帘窗中。\n   - 点击 Play 按钮试听结果。\n   - 调整插件参数优化转录效果，可独立调节源音频和合成转录的音量。\n\n4. **导出 MIDI**\n   满意后，直接将 MIDI 转录结果从插件拖拽至 DAW 的 MIDI 轨道即可完成导出。\n\n> **注意**: NeuralNote 不支持实时转录，因算法需要处理较长的音频块且包含非因果处理步骤。","音乐制作人在 DAW 中录制了一段复杂的即兴钢琴旋律，需要将其转化为可编辑的 MIDI 音符以便后续编曲和配器。\n\n### 没有 NeuralNote 时\n- 手动听写每个音符耗时极长，且容易因听觉疲劳导致音高错误。\n- 面对复调音乐（和弦）时，传统方法难以准确分离同时发声的多个音高。\n- 使用外部转换软件流程繁琐，无法在 DAW 内实时预览调整转换效果。\n- 导出的 MIDI 节奏往往不准，需要花费大量时间手动进行量化修正。\n\n### 使用 NeuralNote 后\n- NeuralNote 直接作为音频插件加载，录音结束瞬间即可在钢琴卷帘窗生成 MIDI。\n- 支持复调转录与弯音检测，准确识别和弦结构及细微的音高变化。\n- 可在播放音频的同时实时调整参数，即时听辨转录结果并优化精度。\n- 内置时间与音高量化功能，满意后直接拖拽即可导出规整的 MIDI 轨道。\n\nNeuralNote 将繁琐的听写工作自动化，让创作者专注于音乐灵感而非技术细节。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FDamRsn_NeuralNote_7c2bd96d.png","DamRsn","Damien Ronssin","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002FDamRsn_f22bd769.jpg",null,"https:\u002F\u002Fgithub.com\u002FDamRsn",[81,85,89,93,97,101,105],{"name":82,"color":83,"percentage":84},"C++","#f34b7d",93.3,{"name":86,"color":87,"percentage":88},"CMake","#DA3434",2.6,{"name":90,"color":91,"percentage":92},"Shell","#89e051",1.5,{"name":94,"color":95,"percentage":96},"Python","#3572A5",1.4,{"name":98,"color":99,"percentage":100},"Batchfile","#C1F12E",0.5,{"name":102,"color":103,"percentage":104},"Inno Setup","#264b99",0.4,{"name":106,"color":107,"percentage":108},"C","#555555",0.3,2522,165,"2026-04-03T08:36:53","Apache-2.0","Windows, macOS, Linux","未说明",{"notes":116,"python":117,"dependencies":118},"Windows 版代码未签名，使用可能需要额外步骤；macOS 版已签名。支持 VST3、AU (仅 Mac) 和独立应用格式。不支持实时转录（因算法延迟）。源码构建需先执行脚本下载 ONNXRuntime 静态库。","3.10 (Windows 源码构建推荐，不支持 3.11)",[119,120,121,122,123],"JUCE","RTNeural","ONNXRuntime","basic-pitch","minimp3",[55,13,53],[126,127,128,129,130,131],"audio","audio-plugin","juce-framework","machine-learning","midi","vst",5,"2026-03-27T02:49:30.150509","2026-04-06T05:15:05.486826",[136,141,146,151,156,161],{"id":137,"question_zh":138,"answer_zh":139,"source_url":140},564,"NeuralNote 是否支持 MIDI 输出流？","是的，NeuralNote 现在支持流式 MIDI 输出。MIDI 输出可以路由到另一个合成器，而不仅仅局限于文件导出。","https:\u002F\u002Fgithub.com\u002FDamRsn\u002FNeuralNote\u002Fissues\u002F41",{"id":142,"question_zh":143,"answer_zh":144,"source_url":145},565,"在 Mac 上无法打开 NeuralNote 独立应用程序怎么办？","默认的 aZip 解压工具可能无法正常工作。尝试使用 `unzip` 命令或其他解压工具重新解压文件。如果问题依旧，请重新下载应用以防文件损坏。","https:\u002F\u002Fgithub.com\u002FDamRsn\u002FNeuralNote\u002Fissues\u002F90",{"id":147,"question_zh":148,"answer_zh":149,"source_url":150},566,"为什么无法加载音频文件？","检查文件名和文件路径中是否包含特殊字符。文件名和路径中不能有特殊字符，否则可能导致加载失败。","https:\u002F\u002Fgithub.com\u002FDamRsn\u002FNeuralNote\u002Fissues\u002F87",{"id":152,"question_zh":153,"answer_zh":154,"source_url":155},567,"NeuralNote 支持哪些音频格式？","目前支持 `.mp3` 和 `.ogg` 格式。如果遇到兼容性问题或在使用旧版本，建议转换为 `.wav`, `.aiff` 或 `.flac` 格式。","https:\u002F\u002Fgithub.com\u002FDamRsn\u002FNeuralNote\u002Fissues\u002F37",{"id":157,"question_zh":158,"answer_zh":159,"source_url":160},568,"如何在 Xcode 中将 NeuralNote 集成到自己的项目中？","需要将以下内容添加到项目中并编译：`Lib\u002FModel` 中的所有代码、ONNX 模型（作为 BinaryData 或修改 `Features.cpp` 从文件加载）、RTNeural、ONNXRuntime（动态库或静态库）。","https:\u002F\u002Fgithub.com\u002FDamRsn\u002FNeuralNote\u002Fissues\u002F84",{"id":162,"question_zh":163,"answer_zh":164,"source_url":160},569,"没有 DAW 如何使用 NeuralNote 转换音频为 MIDI？","你可以构建应用程序（例如 standalone 目标），然后运行应用程序并加载音频文件以获得 MIDI 结果。或者直接安装最新发布的版本使用独立应用。",[166,171,176,181],{"id":167,"version":168,"summary_zh":169,"released_at":170},100194,"v1.1.0","NeuralNote v1.1.0\r\n\r\n# New features \u002F Improvements\r\n- Linux (x86_64) support (VST3 and Standalone).\r\n- Horizontal zoom control on audio and piano roll regions.\r\n- Keyboard shortcuts for play\u002Fpause, reset, record, clear, center and mute buttons.\r\n- Tooltips, with the option to enable\u002Fdisable them.\r\n- Notification when a NeuralNote update is available.\r\n- Updated JUCE\r\n\r\n# Bug fixes\r\n- Fixed Tempo and Time Signature not recalling correctly when loading state.\r\n- Changed `Note Sensibility` and `Split Sensibility` parameters to `Note Sensitivity` and `Split Sensitivity`.","2025-01-11T21:32:54",{"id":172,"version":173,"summary_zh":174,"released_at":175},100195,"v1.0.0","NeuralNote v1.0.0\r\n\r\n## New features \u002F Improvements:\r\n- Installer for both Windows and Mac\r\n  - Windows: includes Standalone and VST3. \r\n  - Mac: includes Standalone, VST3 and AU.\r\n  User chooses what format(s) to install.\r\n- MIDI Output option (transcription can be played with another instrument).\r\n  - Need to take NeuralNote's midi output and route it to another instrument plugin. \r\n  - Not all plugin formats \u002F DAWs support it. Should work with VST3.\r\n  - Tested in Abelton with VST3.\r\n- Parameter management:\r\n  - Parameters are visible to the DAW and can be managed via the DAW.\r\n- State management:\r\n  - Parameters, settings, audio, transcription, etc... are saved on end of session and restored when reloading this session.\r\n  - Recorded files are saved to `\u002FUsers\u002FYourUsername\u002FLibrary\u002FNeuralNote` on Mac and `C:\\Users\\YourUsername\\AppData\\Roaming` on Windows. A recording is deleted when the bin icon is clicked.\r\n- Note Quantization and Time Quantization can now be enabled\u002Fdisabled.\r\n- Time quantization can now be enabled with audio loaded from a file.\r\n  - NeuralNote assumes the start of audio is the start of a measure.\r\n  - Tempo and time signature can be set manually.\r\n- Transcription update on parameter change is significantly faster.\r\n\r\n## Bug fixes:\r\n- Fix incorrect minimum note duration parameter\r\n- Fix flickering playhead\r\n- `.aif` audio files can now be loaded","2024-11-02T23:01:41",{"id":177,"version":178,"summary_zh":179,"released_at":180},100196,"v0.1.0","NeuralNote v0.1.0\r\n\r\nNew functionalities:\r\n- Internal synth to be able to listen to the transcription without exporting MIDI\r\n- Audio player to play the source audio with the transcription (with independent gains)\r\n- No more 3 minutes time limit on recorded and imported audio\r\n- `.mp3` and `.ogg` (Vorbis) file support\r\n- Updated design (icon buttons)\r\n- File browser (thanks @jatinchowdhury18) \r\n- Asio support for Windows Standalone\r\n\r\nBug fixes:\r\n- Fixed incorrect resampling of source audio before transcription\r\n- Fixed pixelated background","2023-09-12T07:15:33",{"id":182,"version":183,"summary_zh":184,"released_at":185},100197,"v0.0.1","NeuralNote v0.0.1","2023-03-24T21:52:48"]